Phat Stellar Cluster Survey. Ii. Andromeda Project Cluster Catalog

Abstract

We construct a stellar cluster catalog for the Panchromatic Hubble Andromeda Treasury (PHAT) survey using image classifications collected from the Andromeda Project citizen science website. We identify 2,753 clusters and 2,270 background galaxies within ∼0.5 deg 2 of PHAT imaging searched, or ∼400 kpc 2 in deprojected area at the distance of the Andromeda galaxy (M31). These identifications result from 1.82 million classifications of ∼20,000 individual images (totaling ∼7 gigapixels) by tens of thousands of volunteers. We show that our crowd-sourced approach, which collects >80 classifications per image, provides a robust, repeatable method of cluster identification. The high spatial resolution Hubble Space Telescope images resolve individual stars in each cluster and are instrumental in the factor of ∼6 increase in the number of clusters known within the survey footprint. We measure integrated photometry in six filter passbands, ranging from the near-UV to the near-IR. PHAT clusters span a range of ∼8 magnitudes in F475W (g-band) luminosity, equivalent to ∼4 decades in cluster mass. We perform catalog completeness analysis using >3000 synthetic cluster simulations to determine robust detection limits and demonstrate that the catalog is 50% complete down to ∼500 M for ages <100 Myr. We include catalogs of clusters, background galaxies, remaining unselected candidates, and synthetic cluster simulations, making all information publicly available to the community. The catalog published here serves as the definitive base data product for PHAT cluster science, providing a census of star clusters in an L spiral galaxy with unmatched sensitivity and quality.

Panchromatic Hubble Andromeda Treasury. XVIII. The High-mass Truncation of the Star Cluster Mass Function

“…Before we proceed with analysis of the cluster mass function, we note that the age distribution of the PHAT young cluster sample is consistent with a near-constant formation history and little or no cluster destruction (Fouesneau et al 2014, Johnson et al 2016. The absence of significant cluster mass loss and destruction over the age and mass range we analyze has an important implication: it is safe to assume that the present day mass function we observe can be interpreted as the initial cluster mass function.…”

“…Cluster masses from MATCH reflect initial masses, unaffected by mass loss from stellar evolution. Cluster ages and masses for the PHAT young cluster sample were published as an appendix in Johnson et al (2016); we publish a full catalog of cluster parameters, demonstrate the reliability of these results using synthetic cluster tests, and compare CMD-based fits to those derived from integrated light SED fitting in A. Seth et al (in preparation).…”

“…We measure the cluster mass function for a wellcharacterized sample of 1249 young star clusters drawn from the PHAT cluster catalog (Johnson et al 2015b). Robust cluster identifications and catalog completeness determinations combine to yield a sample of clusters that is well-suited for a mass function investigation.…”

“…The final sample of clusters was selected according to a candidate's frequency of identification, where each image was examined by >80 AP volunteers. We adopt a cluster identification threshold that maximizes completeness and minimizes contamination with respect to the expert-derived PHAT Year 1 cluster catalog (Johnson et al 2012) and its initial 25% survey coverage.…”

“…Each synthetic cluster was injected into an AP search image and subsequently identified and analyzed in the same way as the genuine clusters; see Section 2.2 in Johnson et al (2015b) for detailed properties of the artificial cluster sample. We compute survey-averaged 50% completeness limits as a function of cluster mass in two bins in cluster age, 10-100 Myr and 100-300 Myr, following a strategy similar to that used in Johnson et al (2016) to account for the spatial variation of completeness and star formation across the survey. First, we bin the synthetic cluster results according to local red giant branch stellar surface density (roughly equivalent to bins of galactocentric radius) to account for the variation of completeness as a function of background stellar density.…”